Use docrep to deduplicate shared parameter docstrings

[endolith]

Summary

Issue #6 suggested using docrep so repeated NumPy-style parameter descriptions are maintained in one place.

This change adds a small docrep dependency and a shared DocstringProcessor in elsim/_docstrings.py, then replaces the duplicated blocks for:

• utilities — the same array_like / 2D utilities wording is reused in strategies (honest_rankings, honest_normed_scores, approval_optimal, vote_for_k) and utility_winner.
• n_voters / n_cands / random_state — shared across random_utilities, impartial_culture, and normal_electorate (the latter uses keep_params slices so dims / corr / disp stay in the middle).

Runtime behavior is unchanged; only __doc__ is assembled at import time. All tests and doctests pass (218 passed).

tests/test_docstrings.py invokes the doc-anchor stub functions so every line of the new anchor code is covered under Codecov, and checks that docrep expanded __doc__ strings include the expected parameter names.

Tradeoffs

• Pros: Single source of truth for repeated parameter prose; matches the approach discussed in Use docrep to simplify docstrings? #6; library is lightweight (depends on six only) and supports Python ≥3.6 per project metadata.
• Cons: One more install dependency; docstrings are no longer fully literal in source (placeholders like %(election_common.parameters)s); contributors need to know the docrep pattern for new reuse.

  

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.49%. Comparing base (8a7213f) to head (1264cce).

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #55      +/-   ##
==========================================
+ Coverage   96.31%   96.49%   +0.17%     
==========================================
  Files          19       20       +1     
  Lines         489      513      +24     
==========================================
+ Hits          471      495      +24     
  Misses         18       18              

Flag	Coverage Δ
no-numba	95.71% <100.00%> (+0.21%)	⬆️
numba	89.66% <100.00%> (+0.50%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.